Collapsing DEL questions

[shaze]

Thanks for Truvari, which we've found very useful.

I do have some questions about place where Truvari doesn't appear to be collapsing variants and I'd like to understand this better,

Here's an example of the truvari_collapsed.vcf file. I've tried various combinations of the collapse parameters and it doesn't seem to make much difference

chr1    109644521       MantaDEL:6979:0:2:0:0:0 A       <DEL>   2258    PASS    END=109648817;SVLEN=4297;SVTYPE=DEL;IMPRECISE;CIPOS=-201,202;CIEND=-314,314;AN=2;AC=1;PctSeqSimilarity=0.8658;PctSizeSimilarity=0.9979;PctRecOverlap=0.9942;SizeDiff=9;StartDistance=-25;EndDistance=-16;TruScore=95;MatchId=77.0       GT:GQ   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.  ./.:.    ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   0/1:47  ./.:.  ./.:.    ./.:.   ./.:.   ./.:.
chr1    109644528       MantaDEL:6883:0:1:0:0:0 G       <DEL>   2258    PASS    END=109648829;SVLEN=4302;SVTYPE=DEL;IMPRECISE;CIPOS=-207,208;CIEND=-217,218;AN=2;AC=1;PctSeqSimilarity=0.8631;PctSizeSimilarity=0.9991;PctRecOverlap=0.9926;SizeDiff=4;StartDistance=-32;EndDistance=-28;TruScore=95;MatchId=77.0       GT:GQ   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.  0/1:60   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.   ./.:.  ./.:.    ./.:.   ./.:.   ./.:.

Any advice or insight would be appreciated.

[ACEnglish]

The problem appears to be that the sequence similarity is turned on even though the ALT sequence isn't reported. I would suggest you separate out your e.g. <DEL> entries and run them without sequence comparison via truvari collapse --pctsim 0.

[shaze]

Many thanks -- that's great.